Table 4.

The performance of machine learning classifiers on AAC based features on main and alternate dataset.

	Main dataset
	Training				Validation
Classifier	Sensitivity	Specificity	Accuracy	AUROC	Sensitivity	Specificity	Accuracy	AUROC
DT	92.269	92.556	92.413	0.962	97.059	91.000	94.059	0.982
RF	95.262	95.533	95.398	0.989	98.039	97.000	97.525	0.994
LR	96.010	96.030	96.020	0.988	98.039	96.000	97.030	0.990
XGB	95.761	95.782	95.771	0.987	98.039	93.000	95.545	0.995
KNN	95.262	95.285	95.274	0.986	97.059	96.000	96.535	0.991
GNB	93.017	98.263	95.647	0.976	93.137	98.000	95.545	0.990
ET	96.010	96.030	96.020	0.991	98.039	96.000	97.030	0.995
SVC	95.761	95.782	95.771	0.987	97.059	96.000	96.535	0.991
	Alternate dataset
	Training				Validation
DT	92.537	92.570	92.557	0.968	94.059	99.379	97.328	0.990
RF	97.015	97.368	97.233	0.995	98.020	97.516	97.710	0.998
LR	96.269	96.285	96.279	0.990	97.030	96.273	96.565	0.987
XGB	97.015	97.059	97.042	0.992	99.010	93.168	95.420	0.998
KNN	95.771	95.975	95.897	0.992	98.020	95.652	96.565	0.995
GNB	92.537	97.059	95.324	0.977	96.040	96.273	96.183	0.983
ET	97.512	97.523	97.519	0.995	98.020	98.137	98.092	0.999
SVC	97.015	96.904	96.947	0.993	98.020	96.894	97.328	0.996

# DT, Decision tress; RF, Random Forest; LR, Logistic regression; XGB, XGBoost; KNN, k-nearest neighbour; GNB, Gaussian naïve base; ET, Extra tree classifier; SVC, support vector classifier.